06-10-2013 12:58 PM
Environment 1:
OS = RHEL 5.3
SFHA 5.0 MP4RP1
Global cluster with replication (two nodes at primary site and one node at DR site)
Environment 2:
OS = RHEL 5.0
SFHA 5.0
Global cluster with replication (two nodes at primary site and one node at DR site)
Both environments are currently running at my two different clients.
As SFHA 5.0 is going to end of support life soon, I want to upgrade SFHA version from 5.0 to 6.0 at both of my clients.
In both cases, I have to install new OS (RHEL >= 5.6). Both clients are running business critical banking applications, So I want to perform the upgrade activity with minimal downtime. The strategy that I have developed for performing this version upgrade is as following:
1. From primary site, remove passive node from cluster membership while leaving application online on active node.
2. Installation of new OS and SFHA on isolated node
3. Mount SAN from live node to the newly installed node (I have already tested that my application works fine with new version of OS and SFHA)
4. Start the application on the newly created node (a new single node cluster)
5. Installation of OS and SFHA on previously active node and run addnode utility to add this node to the single node cluster created in previous step
6. Installation of OS and SFHA on DR node
7. Configuring replication and global cluster option
Anyone please confirm these steps are correct or not. If there is any better strategy, please let me know.
Thanks,
Atif Mehmood Malik
06-11-2013 01:06 AM
By any chance are any of the servers exporting NFS shares here ..if so, some additional handling might be necessary. Other than that - the steps look fine to go ahead.